adding legend to overlaying density plot in ggplot

Question

I have an overlaying density plot from two datasets, created with ggplot2.

  g <- ggplot(data = result_A,
              aes(x = log(cap_universal))) +
    geom_density(fill = "lightblue") +
    geom_density(data = result_B,
                 fill = "purple",
                 alpha = 0.25) +
    xlab("Complete Automation Probability") +
    ylab("Density")

I got what I wanted, a plot looks like this:

However, I have tried many ways but still cannot add legends to this plot. There is no error message, but the legend just won't show.

Will appreciate your help.

Have you tried the `theme` function and its various arguments? You might want to refer to this document: http://www.sthda.com/english/wiki/ggplot2-legend-easy-steps-to-change-the-position-and-the-appearance-of-a-graph-legend-in-r-software — DTYK, Jun 18 '18 at 03:15
@user6606453, have you seen this [post](https://stackoverflow.com/questions/15418302/ggplot2-how-to-show-the-legend) I think it has something to do with your dataset. Can you dput your data? — mnm, Jun 18 '18 at 03:17

Scransom · Accepted Answer · 2018-06-18T06:43:50.343

Combine your data sets result_A and result_B into a single dataset, where the thing these differences are is specified as some factor, e.g. "thing":

result_A$thing <- "A"
result_B$thing <- "B"
result_AB <- rbind(result_A, result_B)

Then instead of calling two datasets with separate calls to geom_density, specify it once with fill = thing and specify your colours and alphas manually, e.g.:

ggplot(data = result_AB, aes(x = log(cap_universal))) + 
  geom_density(aes(fill = thing,
                   alpha = thing)) +
  scale_fill_manual(values = c("light blue", "purple")) +
  scale_alpha_manual(values = c(1, 0.25)) +
  # assuming here that the "result_A" data will plot first as "A" should
  # order before "B" in factors, though can't test w/out yr data
  xlab("Complete Automation Probability") +
  ylab("Density")

This should produce a legend showing what colour your factors "A" and "B" are, which I assume is what you're after.

This is more in-line with the philosophy of ggplot2. However it might also work if you wrap both your fill calls inside aes, e.g. geom_density(aes(fill = "lightblue")). (It's impossible to test this though, because you haven't got a reproducible example above)

Thank you for the idea of combining the two dataset into one. But I was trying to create an overlaying density plot to demonstrate the difference between these two datasets. — user6606453, Jun 18 '18 at 06:01
I've amended the answer so hopefully it "overlays" onto the other data set better for you. Is this what you're after? Note the assumption that your result_A will order first. You might have to switch the factor levels if it comes out the wrong way. — Scransom, Jun 18 '18 at 06:44

adding legend to overlaying density plot in ggplot

1 Answers1